"Lab day," Comet says, dealing 24 blank index cards onto the picnic table.
"Each card gets two coin flips," Wren says. "First flip heads: write Morning. Tails: Evening."
"Second flip heads: write Finished. Tails: Stopped," Comet says. "Then we sort the piles, like the real cards."
Comet flips, Wren writes, Nova tallies. "7 morning and finished. 5 morning and stopped."
"6 evening and finished. 6 evening and stopped," Nova finishes. "Would you like a hint? The second flip never saw the first."
"So the conditional percentages should be close," Wren says. "Finished given morning 58.3%, given evening 50%."
"Close, not equal," Comet says. "Twenty-four cards is a small stack. What do you notice about the real cards?"
"Their gap was much bigger," Wren says. "Bigger than coin flips make by accident."
The crew's counts are made up for the Run Log. Yours will differ, and the comparison is the point.
| Finished | Stopped | Total | |
|---|---|---|---|
| Morning | 7 | 5 | 12 |
| Evening | 6 | 6 | 12 |
| Total | 13 | 11 | 24 |
In the crew's run, finished given morning is 58.3% and finished given evening is 50%. The flips were separate, so the true chances match.
A small stack rarely matches exactly. A gap this small is the kind that coin flips make on their own.
| Statement | True or false? |
|---|---|
| The two coin flips were separate, so the true conditional percentages are equal. | ? |
| 7 + 5 + 6 + 6 = 24 | ? |
| A small stack of cards will always match the true chances exactly. | ? |
| The crew's real sign-up cards showed a gap of 25 points, much bigger than the lab gap. | ? |
Good lab work. Tomorrow you decide what counts as an association, and you use the table as a sample space.